5:15 PM - 6:45 PM
[PPS09-P15] Evaluation of the quantitative estimation of lunar subsurface structures using multi-offset ground penetrating radar
Keywords:Moon, ground penetrating radar, subsurface exploration, FDTD simulation
On the lunar surface, in addition to lava layers with a thickness of several tens meters formed by volcanic activities with flood basalt filling, there are also thin lava layers with a thickness of few meters formed by basaltic plain type volcanism, which is small-scale and intermittent volcanism, and occur after flood basalt filling [Greely, 1975; Gifford and El-Baz, 1981]. The thickness and temporal development of these lava layers have been investigated through observations by optical cameras [Hiesinger et al., 2002; Robinson et al., 2012], subsurface sounder [Oshigami et al., 2014] and ground penetrating radar [Yuan et al., 2021; Feng et al., 2023]. The advantage of subsurface radar observations is ability to detect subsurface lava layers which cannot been found on the ground. However, it is essentially impossible to determine quantitative depth of the subsurface layers and radio wave propagation velocity in the subsurface layers in common offset observation as performed in the previous explorations, in which both transmitting and receiving antennas are mounted on the same rover with keeping their offset distance constant and the rover moves. In such observations, we need to assume some dielectric constant to discuss quantitative thickness of the subsurface layers. On the other hand, among GPR used on the Earth's ground, quantitative estimation of the depth and the propagation velocity of the subsurface layers were achieved by applying multi-offset observation technique in which with the transmitting and receiving antennas placed at different locations with changing offset distance between them. [Daniels, 2004].
The purpose of this study is to evaluate a method for quantitative estimation of the lunar subsurface structures by GPR array observations. We used gprMax [Warren et al., 2016] for FDTD simulations using a model of multiple subsurface 5-meter lava layers with different dielectric constants, and the uppermost 5-meter regolith layer. In the simulation, twenty-one observation points at horizontal intervals of 0.5 m, 1 m, and 2 m were assumed. Using the NMO equation [Dix, 1955], the 2-way travel time (t0) and the RMS average velocity during propagation (Vrms) was obtained from the simulation. From the obtained t0 and Vrms, the buried depth of the lava layer and the dielectric constant in each lava layer were estimated and compared with those in the subsurface structure model used in the simulation. The estimated depths of the subsurface layers in a depth range from 10 to 35 m are consistent with the model (within 10% error) in 1-m and 2-m interval cases. While in 0.5-m interval case, the estimated depths of the subsurface layers in narrower range from 10 to 25 m were consistent with the model, the error of the estimation of the depths of deeper layers were more than 30%. These results suggest that a long array baseline is required to determine the deep subsurface[AK2] structure with high accuracy. As for the estimation of propagation velocity and dielectric constant, the errors were about 10% to 30% and 20 to 50%, respectively, even in lava layers where the depths were successfully estimated within 10% error. This is probably because the error of Vrms becomes larger in deeper layers, and the errors of the velocity and dielectric constant of the deeper layers increased as depth by using Dix equation, the propagation velocity of the n-th subsurface layer is derived from Vrms up to n-th layers and Vrms up to (n-1)-th layers. Velocity structure analysis using the NMO equation is an effective method for quantitative estimation of the depth of the subsurface lava layers, but the application of other methods should be considered for more accurate estimation of the propagation velocity and dielectric constant.
The purpose of this study is to evaluate a method for quantitative estimation of the lunar subsurface structures by GPR array observations. We used gprMax [Warren et al., 2016] for FDTD simulations using a model of multiple subsurface 5-meter lava layers with different dielectric constants, and the uppermost 5-meter regolith layer. In the simulation, twenty-one observation points at horizontal intervals of 0.5 m, 1 m, and 2 m were assumed. Using the NMO equation [Dix, 1955], the 2-way travel time (t0) and the RMS average velocity during propagation (Vrms) was obtained from the simulation. From the obtained t0 and Vrms, the buried depth of the lava layer and the dielectric constant in each lava layer were estimated and compared with those in the subsurface structure model used in the simulation. The estimated depths of the subsurface layers in a depth range from 10 to 35 m are consistent with the model (within 10% error) in 1-m and 2-m interval cases. While in 0.5-m interval case, the estimated depths of the subsurface layers in narrower range from 10 to 25 m were consistent with the model, the error of the estimation of the depths of deeper layers were more than 30%. These results suggest that a long array baseline is required to determine the deep subsurface[AK2] structure with high accuracy. As for the estimation of propagation velocity and dielectric constant, the errors were about 10% to 30% and 20 to 50%, respectively, even in lava layers where the depths were successfully estimated within 10% error. This is probably because the error of Vrms becomes larger in deeper layers, and the errors of the velocity and dielectric constant of the deeper layers increased as depth by using Dix equation, the propagation velocity of the n-th subsurface layer is derived from Vrms up to n-th layers and Vrms up to (n-1)-th layers. Velocity structure analysis using the NMO equation is an effective method for quantitative estimation of the depth of the subsurface lava layers, but the application of other methods should be considered for more accurate estimation of the propagation velocity and dielectric constant.